Format And Type Fonts CCHHEEMMIICCAALL EENNGGIINNEEEERRIINNGG TTRRAANNSSAACCTTIIOONNSS VOL. 39, 2014 A publication of The Italian Association of Chemical Engineering www.aidic.it/cet Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Peng Yen Liew, Jun Yow Yong Copyright © 2014, AIDIC Servizi S.r.l., ISBN 978-88-95608-30-3; ISSN 2283-9216 DOI:10.3303/CET1439003 Please cite this article as: Li B.H., Zhang N., Smith R., 2014, Rate-based modelling of CO2 capture process by reactive absorption with MEA, Chemical Engineering Transactions, 39, 13-18 DOI:10.3303/CET1439003 13 Rate-Based Modelling of CO2 Capture Process by Reactive Absorption with MEA Bao-Hong Li* a , Nan Zhang b , Robin Smith b a Department of Chemical Engineering, Dalian Nationalities University, Dalian 116600, China b Centre for Process Integration, School of Chemical Engineering and Analytical Science, The University of Manchester, Manchester M13 9PL, UK libh@dlnu.edu.cn A rate-based model of CO2 capture process by MEA is built within Aspen Plus V8.0 software in this study. The model is built on the basis of an example model coming along with Aspen Plus V8.0 , thermodynamic model of ENRTL-RK is adopted and both absorption and desorption columns are modelled by rate-based model. Important improvements have been achieved for its accuracy. To be specific, the washing section of the absorption column is strictly modelled and the error on mass balance of MEA is greatly reduced. In addition, the model is validated using the recently published pilot-scale experiment results of the absorption of CO2 by MEA and structured packing Sulzer of mellapak is employed for both absorption and desorption columns in the process. It predicts the experimental profiles of the temperature and the concentration of CO2 in the liquid phase with an accuracy of ±4%, and obviously much better than recently reported model of ±8 %. 1. Introduction Carbon Capture and Storage (CCS) is known as a promising route for controlling the global warming (Kale et al., 2013). Amine scrubbing is probably the only technology for Post-Combustion Capture (PCC) for CO2 that is available to existing power plants (Rochelle, 2009), and most commonly used amine is MEA (an aqueous solution of monoethanolamine). The challenges on the way in implementing the process for treating the entire flue gas of a full size power plant are: how to scale up the amine scrubbing process and how to reduce obviously the exceeding high energy requirement for solvent regeneration (Notz et al., 2012). A recent review of the research work carried out PCC by chemical absorption is given by Wang et al. (2011). To deal with the second challenge, Anantharaman and Berstad (2012) considered those energy integration options in Natural Gas Combined Cycle (NGCC) with PCC, and application of their methodology resulted an improvement in efficiency of the overall process by 0.4 % points. Also, Neveus et al.(2013) considered the interaction between process design and the adopted solvent since the energy efficiency of process modifications depends on the considered solvent. This work focus will be on a reliable and rigorous model which is badly needed to face up to those challenges. Hassan (2005) built an equilibrium-stage type model of the CO2 capture process with MEA using Aspen Plus software. The flue gas data from a cement plant was used and the recommended MEA solvent concentration is 30% by weight to reduce the energy requirement of solvent regeneration to its minimum, the washing section of the absorber column has been neglected in his model to reduce its complexity (Hassan, 2005). Zhang et al.(2009) built a rate-based absorber model for CO2 absorption with MEA and demonstrate the superiority of the rate-based models over the traditional equilibrium-stage models for the pilot plant data from University of Texas at Austin. A rate-based model of CO2 capture process by MEA is built in this study, which is based on an example model within Aspen Plus software (Aspen Technology, 2012). Important improvements have been achieved in accuracy. Specifically, the washing section of the absorption column is strictly modelled by a separate column. In addition, the model is validated using the recently published pilot-scale experiment results of the absorption of CO2 by MEA (Notz et al., 2012). It accurately predicts the experimental profiles mailto:libh@dlnu.edu.cn 14 of the temperature and the concentration of CO2 in the liquid phase, and obviously much better than recently reported model of ±8 % (Kale et al., 2013). Furthermore, this study shows that accuracy rate- based model can be built by selecting suitable unit models and providing proper optional estimates without relying on equation-oriented Aspen Custom Modeler (Kale et al., 2013) and much labour work can be saved. 2. Chemical system and physical properties Since the aim of this work is to demonstrate the rate-based model improvements of a rigorous process model for the reactive absorption CO2 with MEA solvent, the updated electrolyte solution chemistry model coming along with Aspen Plus V8.0 is adopted in this study (Aspen Technology, 2012). Furthermore, NOx and SOx are supposed to be removed completely before they enter the CO2 capture process. A summary of the chemistry model is provided below for completeness and further information can be found in the cited reference (Aspen Technology, 2012). A chemistry model named as MEA is used as the global electrolyte calculation option in the simulation, and the chemical equilibrium is assumed with all the ionic reaction in the MEA chemistry. In addition, two reaction models called Absorber and Stripper have been created. In the Absorber/Stripper models, all reactions are assumed to be in chemical equilibrium except those of CO2 with OH - and CO2 with MEA (Aspen Technology, 2012). 2.1 Chemistry: MEA The chemistry of MEA consists of the following instantaneous reactions: 2 3 MEAH H O MEA H O      (1) 2 3 MEACOO H O MEA HCO      (2) 2 3 2H O H O OH     (3) 2 2 3 3 CO 2H O HCO H O      (4) 2 3 2 3 3 HCO H O CO H O       (5) 2.2 Reaction: Absorber/Stripper The reaction models of Absorber/ Stripper consist of instantaneous reactions (1, 3, 5) and the following finite rate reactions: 2 3 CO OH HCO     (6) 3 2 HCO CO OH     (7) 2 2 3 MEA CO H O MEACOO H O       (8) 3 2 2 MEACOO H O MEA CO H O       (9) The equilibrium constants for reactions 1-5 in MEA chemistry are calculated from the standard Gibbs free energy change. Power law expressions are used for the Rate-controlled reactions (reactions 6-9 in Absorber/ Stripper). Note that Absorber and Stripper share all kinetic parameters except those of reaction 9, those parameters can be found in the reference (Aspen Technology, 2012). The unsymmetric electrolyte NRTL property method (ENRTL-RK) and PC-SAFT equations of state are used to computer liquid and vapour properties, respectively, in this work. CO2, N2 and O2 are selected as Henry-components (solutes) to which Henry’s law are applied. 3. The experiment data The flow diagram of a pilot plant (Notz et al., 2012) is shown in Figure 1. In the process, flue gas is contacted with MEA at ambient pressure in an absorption column and the amine loaded with acidic gases is regenerated in a desorber column. Treated flue gas leaves the absorber column, lean in acid gases. For reducing amine losses, a washing section is installed at the top of the absorber above the lean solvent feed. In the washing section, water is recycled from the liquid collector to the top of the section. The rich solvent leaving the absorber is pumped through a heat exchanger (so called rich-lean heat exchanger) in which it is heated up by the hot lean solvent from the desorber bottom. Then, it is directly distributed on the top of the desorber packing sections. A stripping steam is generated in the desorber bottom by partial evaporation of liquid solvent with electrical heating elements. Above the desorber packing, a washing section similar to that in the absorber is installed to retain amines. Condensate is used as washing water. 15 4. Previous reported simulation 4.1 The simulation model by Kale et al. (2013) The model presented by Kale et al. (2013) consists of a system of differential and algebraic equations, which was implemented in the equation-oriented Aspen Custom Modeler(ACM) simulation environment. To validate the model, the experimental data (Notz et al., 2012) was used, and the comparison of the simulation result with the experimental measurements validates the rate-based model of reactive absorption. The model predicts the experimental profiles of the temperature and the concentration of CO2 in the liquid phase with an accuracy of ±8 %. However, simulation flowsheet is not provided in Kale et al. (2013)’s work and how to simulate the washing section is not explained too. Only the absorption column rather than the absorption-desorption cycle is modeled, the interactions of these two columns have not been explored. Washing section Absorber Desorber Washing water Flue gas Rich solvent Lean solvent Washing section Condensate Electrical heating Treated gas CO2OUT Figure 1: Process flow diagram of absorption-desorption pilot plant 4.2 The simulation model by Aspen Technology, Inc. (2012) Coming with Aspen Plus V8.0, an example rate-based model of CO2 capture process by MEA (Named as Model 1) is developed by Aspen Technology, Inc. in 2012. This example model was just to closely simulate Case 1 of the pilot plant cited in Notz et al. (2012) which incorporates an absorption and a stripper columns, heat exchange between the two columns, and water and amine recycles. The operating conditions of the model were chosen to be as close to those used in Case 1 as possible. An amine makeup stream was not present in the Case 1, but was added to the simulation to maintain mass balance and to aid convergence. The simulation flowsheet of the example model is not provided here because of limited space and can be provided on demand. In this model, the absorber and its washing section are modelled as a complex absorption column of the rate-based type. The whole packing height including the wash section height of 0.42 m and absorber section of 4.2 m is corresponding to 20 stages for simulation, and the washing section is corresponding to stages 1-2 while the absorber section to stages 3-20. The desorber and its washing section are modelled as one column with two packed sections in it. Also, the rate-based model is adopted and 20 stages are specified for the column, Stage 1 represents the condenser at the top, Stage 2 represents the washing section, Stages 3-19 represent the desorber section. And Stage 20 is corresponding to the electrical boiler at the bottom. The simulation results are summarized and compared with those data of the pilot plant in Tables 1-3. Note that a heat exchanger of HX2 is added to change the temperature of the rich solvent to its measured temperature of 112.85 °C before it enters the desorber column. Its heat duty is 262 W, and can be explained as the heat loss during the rich solvent flows from the exchanger HX1 to the stripper column. 16 4.3 The improved rate-based simulation model The simulation flowsheet of the proposed rate-based simulation model (Named as Model 2) is given in Figure 2, it inherits those chemical and reaction models embedded in Model 1 and the following improvements are introduced: (1) The absorber column and its washing section are modelled by two separated columns, one is the absorber of the rate-based type, and the other is the abs-wash of the equilibrium type. The washing section adopts the equilibrium model because there is no obvious benefit when a rate-based one is used. The advantage of this modification is that the function of the washing section can be easily analyzed. (2) Heat loss of the rich solvent after it leaves the absorber column is also considered in Figure 2. The lost heat duty of 389 W reduces its temperature to be 51.66 °C (the same as the measured temperature). Such treatment will make the operating conditions of its subsequent process units closely approach those cited in Notz et al.(2012). Table 1: Comparison of experimental and simulation result- Absorber Experimental Model 1 Model 2 value value error value error Treated gas N2 (g/g) 0.797 0.7980 0.1 % 0.8008 0.5 % O2 (g/g) 0.109 0.1085 0.5 % 0.1088 0.2 % H2O (g/g) 0.072 0.0743 3.2 % 0.0752 4.4 % CO2 (g/g) 0.022 0.0191 13 % 0.0151 31.3 % F (kg/h) 66.6 67.03 0.6 % 66.80 0.3 % T (°C) 47.13 48.21 2.3 % 47.86 1.5 % P (mbar) 977.08 977.08 0 977.08 0 Rich solvent H2O (g/g) 0.661 0.6606 0.1 % 0.6596 0.2 % MEA (g/g) 0.265 0.2658 0.3 % 0.2655 0.2 % CO2 (g/g) 0.074 0.0736 0.5 % 0.0749 1.2 % F (kg/h) 206.5 207.01 0.2 % 207.26 0.4 % T (°C) 51.66 53.35 3.3 % 53.96 4.5 % P (mbar) 1,065.05 987.08 7.3 % 988.2 7.3 % Washing water(in) F (kg/h) 30.48(30.87*) 30.48 0 30.48 0 T (°C) 43.87 43.87 0 43.87 0 Q (W) -61.7 -75 21.6 % -94 52.4 % Note: * marks the original value of 30.87 kg/h is not correct because it does not satisfy the mass balance. 5. Model validation To validate the improved rate-based model, case 1 of pilot plant experiment of the absorption of CO2 with MEA reported by Notz et al. (2012) was used. The main results are summarized and compared in Tables 1-3. In addition, the temperature profile of liquid in the absorber section and the desorber column are provided in Figure 3. From the Figure 3, it is obviously that Model 2 can properly predict the temperature profiles of liquid except the first temperature point of 64.62 °C in the absorber section. Further comparison of Tables 1-3 shows that Model 2 is more closely to the experimental data than Model 1. Note that errors of heat duty are usually larger than 10 %. It can be attributed to heat loss to the environment. 6. Conclusions An improved rate-based model of CO2 capture process is developed in this study. It is validated by the reported pilot plant and compared with an existing example model within Aspen Plus V8.0. Comparison shows that the new model can properly predict Case 1 of the pilot plant and is much more accurate than the example model. Especially, the error on whole MEA mass balance has been reduced from 0.007 kg/h to 0.001 kg/h, and the washing section of the absorber column has been modelled successfully by a separate column. This work shows that accuracy rate-based model can be built by selecting suitable unit models and providing proper optional estimates with Aspen Plus V8.0 software and much labour work can be saved. 7. Acknowledgements We gratefully acknowledge financial support of this work by China Scholarship Council Financial, by CAPSOL project under the Seventh Framework Programme of the European Commission, by Program for 17 Liaoning Excellent Talents in University (LNET) (under Grant No. LJQ2012113) and by Fundamental Research Funds for the Central Universities (under Grant No. DC110104). COOLER Q=-3227 HEATLOS1 Q=-389 RICH-LEA Q=12773 HEATLOS2 Q=922 DESORBER QC=-3866 QR=6554 MAB2 FSPLIT W5 Q=-94 MABS ABS-WASH QC=0 QR=0 PUMP W=18 ABSORBER QC=0 QR=0 56.33 200.115 LEANOUT2 53.96 207.26012 51.66 207.260 12A 51.75 207.260 TOHEATER 111.62 207.260 RICHIN2 112.85 207.260 RICHIN 18.27 5.131 CO2OUT 48.69 2.136 14B 40.16 202.241 LEANIN 48.69 30.666 14 48.69 28.530 14A 46.53 30.480 15-M 43.87 30.480 15 15.00 1.950 WATERMU 50.40 66.981 10-M 47.86 66.795 GASOUT 48.01 72.000 FLUEGAS 18.27 2.019 COND 119.29 200.115 LEANOUT 40.03 200.115 11D 40.03 0.001 MEAMU Temp erat ure (C) M ass Flow Rate (kg/hr) Q Duty (Wat t) W Power(Watt ) Figure 2: The improved simulation flowsheet of CO2 capture process with MEA Table 2: Comparison of experimental and simulation result- Desorber Experimental Model 1 Model 2 value value error value error CO2out H2O (g/g) 0.004 0.0043 7.5 % 0.0044 10 % CO2 (g/g) 0.996 0.9943 0.2 % 0.9942 0.2 % N2 (g/g) 0 0.0011 - 0.0011 - O2 (g/g) 0 0.0003 - 0.0003 - F (kg/h) 4.67* 4.86 - 5.131 - T (°C) 18.27 18 1.5 % 18.27 0 P (mbar) 2,000 2,000 0 2,000 0 Lean solvent H2O (g/g) 0.673 0.6731 0 0.6731 0 MEA (g/g) 0.275 0.2750 0 0.2749 0 CO2 (g/g) 0.052 0.0520 0 0.0520 0 F (kg/h) 200.1 200.09 0 200.11 0 T (°C) 120.8 119.36 1.2 % 119.29 1.2 % P (mbar) - 2,009.1 - 2,005 - Condensate H2O (g/g) 1.0 0.995 0.5 % 0.9930 0.7 % MEA (g/g) 0 0.009 - 0.0006 - CO2 (g/g) 0 0.0041 - 0.0020 - F (kg/h) 2.04 2.07 1.5 % 2.02 1 % T (°C) 16.02 18 11.1 % 18.27 14 % P (mbar) - 1,999.1 - 2,000 - Condenser duty (W) -2,773.6 -3,257 17.4 % -3,866 39.4 % Reboiler duty (W) 7,951.5 7,050 11.3 % 6,554 17.6 % Note: * marks the value should be questioned because the whole mass balance is not satisfied if the flowrate of other input and output streams are corrected. Specifically, for water balance in Figure 2, the inlet streams are Watermu and Fluegas while the outlet streams are Gasout, CO2out and Cond. 18 Table 3: Comparison of experimental and simulation result- key process parameters Experimental Model 1 Model 2 value value Error(abs) value Error(abs) Water makeup F (kg/h) 1.95 1.949 0.001 1.950 0 MEA makeup F (kg/h) 0 0.007 0.007 0.001 0.001 Rich-Lean Exchanger duty (W) 13,524 12,424 1,100 12,773 751 Cooler on lean solvent duty (W) -2,866.9 -3,593 726.1 -3,364 497.1 Fluegas F (kg/h) 72.0 72.0 0 72.0 0 (a) (b) Figure 3: The temperature profile of liquid in (a) the absorber section, (b) the desorber column(x axis represents stages while y axis represents temperature) References Anantharaman R, Berstad, D, 2012, Energy integration in an NGCC plant with post combustion CO2 capture-systematic methodology for evaluating process alternatives. Chemical Engineering Transactions, 29, 451-456, DOI: 10.3303/CET1229076. Aspen Technology, Inc., 2012, Rate-based Model of the CO2 Capture Process by MEA using Aspen Plus. Burlington, USA. Hassan S. N., 2005, Techno-economic study of CO2 capture process for cement plants. University of Waterloo Ontario, Canada. Kale C., Gorak A., Schoenmakers H., 2013, Modelling of the reactive absorption of CO2 using mono- ethanolamine. International Journal of Greenhouse Gas Control, 17, 294-308. Neveux T., Moullec Y. L., Corriou J. P., Favre E. 2013, Energy performance of CO2 capture processes: interaction between process design and solvent. Chemical Engineering Transactions, 35, 337-342, DOI:10.3303/CET1335056. Notz R., Mangalapally H.P., Hasse H., 2012, Post combustion CO2 capture by reactive absorption: Pilot plant description and results of systematic studies with MEA. International Journal of Greenhouse Gas Control, 6, 84-112. Rochelle G.T., 2009, Amine Scrubbing for CO2 Capture. Science, 325, 1652-1654. Wang M., Lawal A., Stephenson P., Sidders J., Ramshaw C., 2011, Post-combustion CO2 capture with chemical absorption: a state-of-the-art review. Chem. Eng. Res. Des., 89, 1009-1624. Zhang Y., Chen H., Chen C.C., Plaza J. M., Dugus R., Rochelle G.T., 2009, Rate-based process modelling study of CO2 capture with aquous monoethanolamine solution. Ind. Eng. Chem. Res., 48, 9233-9246. http://apps.webofknowledge.com/OneClickSearch.do?product=WOS&search_mode=OneClickSearch&excludeEventConfig=ExcludeIfFromFullRecPage&colName=WOS&SID=Y2TgKp9ddLVE7MqpsMh&field=AU&value=Anantharaman,%20R http://apps.webofknowledge.com/OneClickSearch.do?product=WOS&search_mode=OneClickSearch&excludeEventConfig=ExcludeIfFromFullRecPage&colName=WOS&SID=Y2TgKp9ddLVE7MqpsMh&field=AU&value=Berstad,%20D